Hacking Hardware Picofly - a HWFLY switch modchip

[abal1000x]

.

 

[fox87]

Hey guys. If the tin just wont stick to the APU's GRD / Source pad, what am i doing wrong?

 

[abal1000x]

.

 

[fox87]

Easier to just use the shield as GND.

Yeah. I did that as Plan b now. But would like to advance, to learn how other people do it.

 

[poiu15]

Is there an alternative pad for 3.3v on oled board? Just don't like to solder on components.

 

[jkyoho]

Is there an alternative pad for 3.3v on oled board? Just don't like to solder on components.

 

[poiu15]

View attachment 378874

Got it, thank you!

 

[Adran_Marit]

Hello guys. Is this the cheapest method to backup, restore boot0, boot1 and raw of oled emummc using linux?
https://www.aliexpress.com/item/
I also found this but it is out of stock and it was much cheaper
https://www.tindie.com

Question... how ya getting the emmc off the oled to clip onto the adaptor? ;P

this is good for v1/v2 switches but not so much lites or oleds

 

[LogicalMadness]

@Switxh I was able to fix the 'slow gitch' on my Lite.

Storytime. Completed a Lite (now I know it's a Samsung) for a client that had 0.1mm wire, PZero-2.67FW and a flex cable installed. Glitch times seemed fine, from what I stoner-remember, but I do also remember thinking they weren't great\consistent. They still haven't picked up the Lite and I completed it weeks ago, so I stole it's PZero for another client while I waited for my second batch of more chips.

The new chip I eventually installed had 2.74FW and glitch times seemed like crap. I decided to ditch the garbage flex and replaced it with a single-FET-single-cap-install using my new 32AWG wire. Same slow glitching though. SOMETIMES it would fast glitch twice in a row and I would think 'got it!', but then 30-60 sec (idk exactly, didn't time) glitch again.

Downgraded to 2.7FW, same issue. 2.67FW, same issue. Fk it, all new wires, same issue. Add another FET, works! Update to 2.74FW, works instantly, CONSISTENTLY!

I don't remember everything you tried, but I hope that may help you\others.

 

[FlowerFire]

Hey guys, I've got this install where I can run both emunand and sysnand, however, if I run complete sysnand backup, I get the "slow emmc" error. Sometimes with both sysnand and ofw I had a crash error (but forgot to write down the error code).
I guess the two things are related, right? Is this one of those cases where i need to put more resistors than the default three?

 

[Switxh]

@Switxh I was able to fix the 'slow gitch' on my Lite.

Storytime. Completed a Lite (now I know it's a Samsung) for a client that had 0.1mm wire, PZero-2.67FW and a flex cable installed. Glitch times seemed fine, from what I stoner-remember, but I do also remember thinking they weren't great\consistent. They still haven't picked up the Lite and I completed it weeks ago, so I stole it's PZero for another client while I waited for my second batch of more chips.

The new chip I eventually installed had 2.74FW and glitch times seemed like crap. I decided to ditch the garbage flex and replaced it with a single-FET-single-cap-install using my new 32AWG wire. Same slow glitching though. SOMETIMES it would fast glitch twice in a row and I would think 'got it!', but then 30-60 sec (idk exactly, didn't time) glitch again.

Downgraded to 2.7FW, same issue. 2.67FW, same issue. Fk it, all new wires, same issue. Add another FET, works! Update to 2.74FW, works instantly, CONSISTENTLY!

I don't remember everything you tried, but I hope that may help you\others.

View attachment 378890

Wow yes this sounds exactly like the issue I have with an OLED! Especially when you said about how you would sometimes get a couple of fast glitches in a row and think "it's fixed" then the next glitch takes absolutely ages LOL! I fitted mosfets but not in the way you have with it on a single cap. Do you have a diagram of how to wire them the way you have done it? Thanks!

 

[QuiTim]

Hey guys, I've got this install where I can run both emunand and sysnand, however, if I run complete sysnand backup, I get the "slow emmc" error. Sometimes with both sysnand and ofw I had a crash error (but forgot to write down the error code).
I guess the two things are related, right? Is this one of those cases where i need to put more resistors than the default three?

Yep, that's the case most likely. You have a section on this in the Definitive Guide

 

[malgamer]

i got 2 oled, one with dat0 resistor around 600 and it boot within a sec, another one with black adapter and much higher 800 ohm i guess. This second one struggle with data0 error, after some reflow, the resistor drop a bit but still way slower than the first one

maybe yours second unit oled emmc was skhynix brand

 

[LogicalMadness]

Wow yes this sounds exactly like the issue I have with an OLED! Especially when you said about how you would sometimes get a couple of fast glitches in a row and think "it's fixed" then the next glitch takes absolutely ages LOL! I fitted mosfets but not in the way you have with it on a single cap. Do you have a diagram of how to wire them the way you have done it? Thanks!

 

[meeebro]

How about cutting on the red line? So he won't need to find the new position, and if the connection has no problem, maybe the big fat right side stuck between dat0 and NC pin will make the contact more stable. View attachment 378867

I have the crappy OAT0 adapter, tried cutting the red area first, still get 0.720 V and the error. I cut the left as well and a little of the other corner to move it more right and I still get 0.72v and the error. At least the switch boots into stock HOS still. I guess I will wait for a new adapter, is there a link to better ones?

 

[johndoenut]

Use this point shown by mouse cursor.
Tested, but this point is no joke and you need uvmask or glue to secure the wire. I ve done probably 5 lites with this point, also its now my preferred dat0 point onwards.

Opps! sorry for not clarifying. The pad that I repaired was cmd. Thanks for the dat0 point though, I'll keep it in mind.

Have you ever used this other point for cmd though in the included picture? I used the pcb scan to follow the trace backwards (or forwards, lol) . Do you think it would be suitable? I superimposed the picture of the trace so the path could be clearer.

 

[jkyoho]

Opps! sorry for not clarifying. The pad that I repaired was cmd. Thanks for the dat0 point though, I'll keep it in mind.

Have you ever used this other point for cmd though in the included picture? I used the pcb scan to follow the trace backwards (or forwards, lol) . Do you think it would be suitable? I superimposed the picture of the trace so the path could be clearer.View attachment 378911

You need to make the ripped LEFT pad connecting back to the right pad, no point to make a long wire to the resistor and RIGHT pad not damage so resistor to the RIGHT pad is in good condition.

 

[johndoenut]

You need to make the ripped LEFT pad connecting back to the right pad, no point to make a long wire to the resistor and RIGHT pad not damage so resistor to the RIGHT pad is in good condition.

As I mentioned in my other posts earlier in the thread, I already fixed that but I didn’t get pictures of the repair. What I’m referring to is using the far alternate point so I won’t disturb the bridge I made. Or worse, rip the pad that’s left.

 

[cgtchy0412]

Opps! sorry for not clarifying. The pad that I repaired was cmd. Thanks for the dat0 point though, I'll keep it in mind.

Have you ever used this other point for cmd though in the included picture? I used the pcb scan to follow the trace backwards (or forwards, lol) . Do you think it would be suitable? I superimposed the picture of the trace so the path could be clearer.View attachment 378911

Warning.... dont use that alter cmd point.. its very very easy to ripped out and result to purple screen .. not worth the headache.
Just use normal point.. its easy with enough cleareance.
Also if you install pico on top of ram module you can use this big fellow ground point (tested).

 

[jkyoho]

As I mentioned in my other posts earlier in the thread, I already fixed that but I didn’t get pictures of the repair. What I’m referring to is using the far alternate point so I won’t disturb the bridge I made. Or worse, rip the pad that’s left.

I see, A broken resistor is easy repair compared to a ripped pad. YMMV